Detection and mitigation of ionospheric stripes in PALSAR data

Due to its lower frequency, the ALOS PALSAR L-Band synthetic aperture radar (SAR) is more susceptible to ionospheric effects than the higher frequency C-Band (e.g., RADARSAT-2 and Envisat ASAR) and X-Band (e.g., TerraSAR-X and COSMO-SkyMED) space-based SAR systems. During its 5 years in operation, PALSAR collected more than 2 million scenes, and ionospheric effects were observed in scenes from the equatorial region to the polar regions [1, 2, 3]. Anomalous stripes from ionospheric disturbances are present in thousands of PALSAR scenes, primarily near the magnetic equator [1]. The current method of detecting the ionospheric stripes is by visual inspection of the scenes, which is extremely challenging for 2 million scenes. These ionospheric stripes degrade image quality, interferometric products, and change detection maps [2]. ALOS-2 is expected to be launched in 2013. Since this SAR system will have more than twice the frequency bandwidth of PALSAR and the ionospheric activity will be greater during its period of operation, more ALOS-2 scenes will be affected by the ionosphere. This paper presents a method to detect and mitigate the ionospheric stripes in PALSAR data using spectral filters in the frequency domain. This method has the potential of being fully automated, which may significantly reduce the search time for scenes with ionospheric stripes.